基于VECM和BLCOP模型的动态行业战术性资产配置

Dynamic Industry Tactical Asset Allocation Based on VECM and BLCOP Models

选用上证行业指数中5个表现较好、相关性较为稳定的行业指数从2009年1月到2019年2月的月度数据进行动态建模.运用向量误差修正模型对收盘价进行拟合,并把预期收益作为绝对观点的收益;采用多维有偏t分布作为收益的先验分布,并把观点与先验结合导出具有t-Copula相依性结构的Black-Litterman(BLCOP)模型.实证分析显示在最大夏普比策略下BLCOP模型较等权重、传统BL模型及其先验分布有更高的收益,且基于BLCOP模型的权重分配较其先验有明显变化,说明应用BLCOP模型进行动态配置是有必要的.最后,分析BLCOP模型在最大化收益风险比策略下采用3种风险指标时的组合收益差异,以帮助不同风险偏好的投资者提高投资组合战略价值.

Monthly data of 5 industry indexes in Shanghai are taken for dynamic modeling,which havegood performance and stable correlation from Jan.2009 to Feb.2019.Vector error correct model is applied to fit the closing price data,then predict returns are taken as absolute opinion return.Multidimensional biased t-distribution is employed to fit the return data and taken as prior distribution,a Black-Litterman model with t-Copula dependence structure(BLCOP)is derived by combing opinion and prior.The empirical results show BLCOP model has higher return than equal-weight,traditional Black-Litterman model and their priors under the strategy of maximizing Sharpe Ratio, and the weightallocation based on BLCOP model has significant difference from the case based on its prior,whichindicates it is meaningful to make dynamic allocation based on BLCOP model. Finally,the returndifference among portfolios is analyzed under the strategy of maximizing the return-risk ratio with threerisk indicators to help investors with various risk preference to improve portfolio strategic value.